Fingerprint Clustering

The fingerprint clustering algorithm used is the Jarvis-Patrick clustering method. Two molecules are considered to be in the same cluster if the lists of molecules to which they are similar overlap to a reasonable degree. In other words, two molecules belong in the same cluster if they are similar to the same set of molecules. The Jarvis-Patrick procedure is O(n2), where n is the number of items to cluster; therefore, a substantial amount of time is required if n is large.

Options

Molecule Field
Field(s) containing the molecules
Fingerprint
Select fingerprint that should be used for clustering
Fingerprint (other)
Name of the other fingerprint.
Metric
Select similarity metric. Count tanimoto is only supported for ECFP and FCFP fingerprints. tanimoto-ss is not supported for ECFP and FCFP.
Method
Select clustering method
Similarity
The percentage similarity that is used to determine whether two molecules are similar. (Fingerprints)
Overlap
The percentage overlap that is used to determine whether two molecules belong to the same cluster. (Neighbor Lists)
Number of Clusters
Enter the number of clusters
Batch Size
Size of batches for mini-batch K-Means
Number of Outliers
Enter the number of outliers omitted for centroid calculation

Input Ports

Icon
DataTable

Output Ports

Icon
Clustered DataTable

Views

Fingerprint Clustering
Summary

Workflows

  • No workflows found

Links

Developers

You want to see the source code for this node? Click the following button and we’ll use our super-powers to find it for you.